Plasma noradrenaline is little used in evaluating the prognosis of cardiac failure because of the theoretical necessity of interrupting treatment for a few days before blood sampling. The present study reevaluated the prognostic value of this parameter with blood sampling performed during treatment and then 48 hours after withdrawal of treatment in 192 patients with chronic stable cardiac failure at an advanced stage (64% of patients in Classes III or IV with an average ejection fraction of 28.5 +/- 13.5%). During follow-up (average 43 months) there were 51 deaths and 17 transplants. None of the patients were lost to follow-up. Univariate analysis of 52 variable observers during the initial phase of evaluation found in decreasing order of predictive value for death plasma noradrenaline levels before and after withdrawal of treatment for 48 hours. Serum sodium, age, systolic mean and diastolic pulmonary artery pressures. In multivariate analysis: noradrenaline with or without withdrawal of treatment, hyponatraemia and systolic pulmonary artery pressure. Actuarial survival curves distinguished the following parameters: noradrenaline levels became predictive at concentrations of over 210 pg/mL and there was a significant difference in survival with respect to 4 levels of serum noradrenaline (with or without treatment) > 300 pg/mL, 300 to 600 pg/mL and > 900 pg/mL. This serum noradrenaline measured without withdrawal of treatment (especially angiotensin converting enzyme inhibitors) is a powerful predictor of mortality, carrying a progressively poorer prognosis as the concentration increases.